Content uploaded by Dan Roberts
Author content
All content in this area was uploaded by Dan Roberts on Jul 17, 2014
Content may be subject to copyright.
MARINE ECOLOGY PROGRESS SERIES
Mar Ecol Prog Ser
Vol. 321: 19–30, 2006
Published September 8
INTRODUCTION
Scales of patchiness and variation in assemblages,
either in time or space, have become an important
focus of many marine ecological studies (Underwood &
Chapman 1998a). They have 2 key foci. First, studies
that determine the spatial and temporal scales at
which species and assemblages show most variability,
allow one to identify the scales at which important eco-
logical processes are likely to be operating (e.g.
Archambault & Bourget 1996, Thompson et al. 1996,
Underwood & Chapman 1998b, Hewitt et al. 2002) and
the scales at which disturbance and recovery of assem-
blages are likely to be shown (Warwick & Clarke 1993,
© Inter-Research 2006 · www.int-res.com
*Email: dan@bioanalysis.com.au
Structure and dynamics of sponge-dominated
assemblages on exposed and sheltered temperate reefs
D. E. Roberts
1, 2, 3,
*
, S. P. Cummins
3
, A. R. Davis
2
, M. G. Chapman
3
1
BIO-ANALYSIS: Marine, Estuarine & Freshwater Ecology, 7 Berrys Head Road, Narara, New South Wales 2250, Australia
2
Institute for Conservation Biology, School of Biological Sciences, University of Wollongong, Wollongong,
New South Wales 2522, Australia
3
Centre for Research on Ecological Impacts of Coastal Cities, Marine Ecology Laboratories A11,
University of Sydney, Sydney, New South Wales 2006, Australia
ABSTRACT: There have been few studies on the structure and dynamics of sponge-dominated
assemblages, despite the fact that such assemblages are vulnerable to environmental impacts from
many anthropogenic disturbances. Sponges are generally slow to recruit, slow growing and long
lived; hence, they may be very vulnerable to anthropogenic and natural disturbances. In order to
understand how such assemblages may respond to disturbance, it is essential to measure natural pat-
terns of spatial differences and temporal changes, so that any future impact assessments can be iden-
tified. This study quantified and contrasted patterns of abundance in sponge-dominated assemblages
on deep reefs (18 to 20 m) exposed to direct oceanic swell with reefs in the more protected entrances
to large embayments. We examined the hypothesis that erect sponges would dominate the reefs
within embayments, while encrusting species would be more prevalent on the wave-exposed reefs.
We also predicted that wave-exposed reefs would show greater temporal and spatial variability. Four
reefs within embayments and 4 open coastal reefs, each with 3 nested sites, were sampled with ran-
domly placed photo-quadrats. Sponges dominated the reefs we examined, accounting for around
25% of the cover of the substratum on exposed reefs and usually >40% on sheltered reefs. A total of
82 species of sponge were identified. As predicted, erect sponges accounted for the majority of the
species richness and cover of sponges on sheltered reefs, whereas encrusting species predominated
on the exposed reefs. The contribution of other sessile invertebrates to the cover and species richness
on these reefs was small. nMDS plots revealed striking and consistent differences in the assemblages
between the exposed and sheltered reefs, although PERMANOVA failed to detect significant differ-
ences. ANOVA revealed significant fluctuations in the cover and richness of taxa at various spatial
and temporal scales. Examination of the components of variation of selected taxa showed that most of
the variability was found at the smallest spatial scale, i.e. within the residual, and this variability was
generally greatest for taxa on exposed reefs compared with sheltered reefs. These results have
important implications for monitoring programmes designed to detect environmental impacts on
sponge-dominated assemblages.
KEY WORDS: Porifera · Marine assemblages · Ecological patterns · Wave exposure · Sessile
invertebrates · Scales of variation
Resale or republication not permitted without written consent of the publisher
Mar Ecol Prog Ser 321: 19–30, 2006
Bishop et al. 2002). The second key focus relates to the
scales that must be incorporated into sampling strate-
gies needed to identify natural or anthropogenically
induced changes to assemblages (Underwood 2000,
Stewart-Oaten & Bence 2001).
Previous studies show that many marine assem-
blages vary at a hierarchy of scales, from centimetres
to 100s of kilometres and from days to years (Coleman
2002, Hewitt et al. 2002, Davis et al. 2003). Most of
these types of studies have, however, been done in the
intertidal or in relatively shallow water, and, conse-
quently, our understanding of the spatial and temporal
scales of variability in invertebrate assemblages in
deep water is scant and cannot be readily predicted
from previous work.
Deeper subtidal reefs in temperate waters, such as
those below 18 m on the east coast of Australia, are
generally dominated by sessile invertebrates, com-
prised predominantly of sponges, but also ascidians,
bryozoans and cnidarians (Roberts & Davis 1996,
Hooper & Kennedy 2002). Sponges are renowned for
their morphological plasticity (Hill & Hill 2002), partic-
ularly in relation to wave action or flow (Palumbi 1986,
Barthel 1991, Ginn et al. 2000, McDonald et al. 2003).
Encrusting or prostrate forms closely adhere to the
substratum and are considered to be better suited to
high-energy environments, such as those found on
shallow subtidal reefs (Roberts & Davis 1996, McDon-
ald et al. 2002). In contrast, erect or massive forms pos-
sessing a relatively small basal area relative to volume
do poorly in such environments (Wulff 1995, Bell &
Barnes 2000), while they often dominate deeper tem-
perate reefs (e.g. Roberts & Davis 1996). Given their
prevalence on deep reefs and their longevity, sponges
are excellent organisms with which to explore predic-
tions about the size and scale of disturbances.
The structures of invertebrate assemblages on shal-
low temperate reefs are modified by physical and
biological disturbances (Ayling 1981, Underwood et al.
1991), and it appears that these processes occur at
relatively small scales (<10 m; Davis et al. 2003).
Deeper, subtidal reefs also routinely experience sig-
nificant exposure to water movement from storms and
ocean waves (Short & Trenaman 1992), which may
play an important structuring role at larger spatial
scales. It has been argued that natural physical
processes (e.g. storms) may account for a large pro-
portion of the variation in assemblages observed
between habitats on exposed coastlines (Underwood
et al. 1991, Wulff 1995, Posey et al. 1996, Underwood
1998, 1999), whereas, in more sheltered locations, e.g.
the subtidal reefs at the entrances to estuaries and
embayments, the frequency and severity of the phy-
sical effects of exposure will be greatly reduced. In
these habitats, strong tidal currents or episodic pulses
of freshwater after heavy and prolonged rainfall may
play a significant role in some of the processes deter-
mining the structure and dynamics of sponge-domi-
nated assemblages (Hummel et al. 1988).
Here, we examine the spatial and temporal variabil-
ity of sponge-dominated assemblages on exposed and
sheltered reefs at depths of 18 to 20 m, in temperate
New South Wales, Australia. Much of our understand-
ing of these assemblages stems from studies on artifi-
cial substrata or natural reefs in shallow water (e.g.
Ayling 1981, Kay & Butler 1983, Chapman et al. 1995,
Butler & Connolly 1996). Predictions from such studies
may not hold on natural surfaces in deeper (>18 m
depth) water. Importantly, because of large differences
in physical conditions between sheltered and exposed
reefs, e.g. water movement (Palumbi 1986, Barthel
1991, Graham et al. 1997), siltation (Bell & Barnes
2000), potential differences in recruitment (Maldonado
& Young 1996, Maldonado & Uriz 1998), competition
and predation (Wright et al. 1997), one can propose
that patchiness in these assemblages, at scales of
metres to 10s or 100s of metres, may differ between
sponge-dominated assemblages on sheltered and
those on exposed reefs. Quantification of such differ-
ences would therefore focus attention on the scales at
which processes may differ between these habitats.
These are important considerations given that wave-
exposed reefs are targeted for the disposal of sewage
(Roberts et al. 1998), while sheltered reefs are often
subject to urbanisation and the development of port
facilities (Carballo et al. 1996).
Specifically, we tested the hypothesis that spatial
and temporal patterns of variability in sponge-domi-
nated assemblages would be different on exposed
reefs compared with those on sheltered reefs at the
same depth. We believe that storm-related physical
disturbance will lead to elevated dynamism and vari-
ability in wave-exposed assemblages by releasing
space for settlement and recruitment. We predicted
that: (1) assemblages on wave-exposed reefs would
show spatial and temporal variation at larger scales,
but possess lower sponge cover than assemblages on
sheltered reefs; (2) sheltered reefs would have more
species and cover of erect sponges compared with
sponges on exposed reefs; (3) exposed reefs would
have more species and cover of encrusting sponges
relative to sheltered reefs.
MATERIALS AND METHODS
Study locations and sampling. Spatial and temporal
patterns in subtidal, sponge-dominated assemblages
were determined by sampling exposed and sheltered
reefs, between Botany Bay and Broken Bay, New
20
Roberts et al.: Sponge assemblages on temperate reefs
South Wales, Australia (Fig. 1); 4 exposed and 4 shel-
tered estuarine reef locations were sampled at depths
of approximately 18 to 20 m. Some of these reefs have
been described previously (see Underwood et al. 1991,
Chapman et al. 1995, Roberts & Davis 1996), but no
previous study has incorporated the spatial and tempo-
ral scales of the present study to quantify the relative
scales of variability for the suite of sponges in these
habitats.
The assemblages on these reefs were sampled
using a diver-operated camera rig that supported a
35 mm Sea & Sea Motor Marine-2 underwater cam-
era and strobe. The assemblages at each location
were sampled on 7 random occasions from April
1993 to September 1995. Within-location variability
in the assemblages was determined by haphazardly
photographing 5 replicate quadrats (photo-quadrat
dimensions: 0.8 m × 0.56 m, total area 0.45 m
2
) at
each of 3 randomly nested sites (approximately 50 m
in diameter and 50 m apart) within each location.
Within each location, different sites were sampled
at each time.
The photo-quadrats were analysed using a Bell and
Howell ‘black box’ projector. An overlay plastic grid of
100 regularly spaced points was placed on the screen,
and estimates of the percentage cover and number of
species were recorded from the photo-quadrat. Many
of the crustose coralline and foliose macroalgae could
not be identified to species. They were grouped into
morpho-taxa and termed foliose macroalgae and crus-
tose coralline algae. To help differentiate the taxa
recorded in photo-quadrats, specimens were collected
at all locations. An in situ, close-up, 35 mm colour pho-
tograph was taken of each specimen prior to collection,
as a permanent record of the habit of the organism.
Many invertebrates (especially sponges) lose colour
and shape once out of the water, so another photo-
graph was taken on the surface and the samples were
labelled and immediately frozen for later identifica-
tion. This voucher collection is lodged with the
Queensland Museum, Australia.
Data analysis. Differences in the
assemblages between exposed and
sheltered locations were tested using a
nested design (locations nested in
exposure and sites nested in locations)
using PERMANOVA (=NPMANOVA;
Anderson 2001) on Bray-Curtis dissimi-
larities calculated from untransformed
data. Analyses were done for each time
separately, to provide 7 independent
estimates of variability for each spa-
tial scale. These analyses identified
whether there was significant variation
among sites and locations, as an esti-
mate of relative variability at different
scales, although pairwise tests were not
done to examine this further, as sites
and locations are random factors,
selected to represent spatial scales at
which one might expect these assem-
blages to vary. Non-metric multi-
dimensional scaling (nMDS) ordina-
tions were used to graphically illustrate
patterns in the assemblages.
To test the hypotheses that the
amount of small-scale variability in
these assemblages at the scale of
metres (among replicates within a site)
and 10s of metres (among replicates
between sites) was consistent among
locations and consistently different
between sheltered and exposed condi-
tions, Bray-Curtis dissimilarities were
calculated between randomly paired
21
Fig. 1. Locations of the reefs in exposed (Bungan Head, Long Reef, North
Head, Cape Banks) and sheltered (IP: Inscription Point; HH: Henry Head;
QH: Quarantine Head; SH: South Head) habitats along the Sydney coastline
Mar Ecol Prog Ser 321: 19–30, 2006
replicates within or between sites at each time of sam-
pling. These measures can be used in an analysis of
variance framework, because each replicate measure
is an independent estimate of variability (Underwood
& Chapman 1998a). These data do not test whether
variability in the assemblage increases as spatial scale
increases (as per PERMANOVA and ANOVA for mul-
tivariate and univariate measures, respectively), but
specifically test the hypothesis that the magnitude of
small-scale variability is consistent at larger scales.
This will indicate whether small-scale variability is
simply ecological ‘noise’, or whether there are small-
scale processes that consistently act over larger scales.
These measures were compared between exposures
and among locations (and sites for measures of within-
site variability) using analyses of variance, with times
(random), exposure (fixed), locations and sites (random
and nested).
To measure spatial variation within sites, 2 pairs of
replicates per site were paired, to provide 2 indepen-
dent Bray-Curtis dissimilarities per site. To measure
variation at the scale of 100s of metres, replicates were
paired between sites (2 replicates in Site 1 with 2 in
Site 2, 2 in Site 1 with 2 in Site 3 and 2 in Site 2 with 2
in Site 3), giving 6 independent Bray-Curtis dissimilar-
ity measures per location per time.
Temporal variation in the assemblage was measured
for each site by calculating the average Bray-Curtis
dissimilarities among times using the site centroids.
These were then similarly compared between expo-
sure and among locations, using the temporal variation
within each site as a replicate measure.
Prior to ANOVA, data for selected taxa were exam-
ined for homogeneity of variances using Cochran’s test
(Winer 1971). Where variances were heterogeneous,
data were ln(x + 1) transformed for number of taxa and
arcsine transformed for percentage cover (Winer
1971). Where transformations did not result in homo-
geneous variances, analyses were done on the
untransformed data (Underwood 1997). If variances
could not be stabilised at p = 0.05, but could be sta-
bilised at p = 0.01, ANOVA was interpreted using the
p = 0.01 probability level (Underwood 1997).
In addition, the spatial scales at which most variation
could be explained were examined for the cover of
selected taxa. The relative contribution of each scale to
the total variation was calculated from the components
of variation. These were estimated from mean squares
using the untransformed data in the ANOVA (Under-
wood 1997). When negative estimates of components
of variation were found, that source of variation was
removed from the model and all components recalcu-
lated using the method described by Fletcher & Under-
wood (2002). Two-tailed F-tests were used to compare
the residual variation between exposed and sheltered
reefs (Underwood 1997). Comparisons of variance
components for sites and locations using 2-tailed F-
tests are not valid, and these results were interpreted
qualitatively.
RESULTS
Sponges were the most diverse and abundant taxon
encountered on sheltered and on exposed reefs. A total
of 82 species of sponges were identified. Of the other
major phyla, 14 ascidians, 12 bryozoans and 12 cnidar-
ians were also identified. The most abundant algae
were crustose Corallinacea and a mixture of macro-
scopic foliose species. A matrix of silt, consisting of a
mixture of micro-flora and fauna, silt and micro-organ-
isms, was the dominant primary cover on exposed
reefs and, at some times, the dominant cover on shel-
tered reefs.
Patterns in assemblage structure
PERMANOVA showed a significant difference
among replicate sites on 4 of 7 occasions and among
locations on all occasions. Therefore, data could not be
pooled across sites or locations to increase the power of
the test for the main effect of exposure. A significant
difference was only determined between exposures in
April 1993 (F
1,6
= 3.02, p < 0.05). Nevertheless, nMDS
ordinations clearly plotted assemblages on exposed
reefs as separate from those on sheltered reefs in April
1993, April 1994, January 1995 and April 1995, indicat-
ing that these 2 conditions do support distinctly differ-
ent assemblages (Fig. 2). In August 1993 and January
1994, only locations at North Head overlapped with
those in the sheltered locations, whereas, in Septem-
ber 1995, assemblages were much more similar.
When the data for all exposed or sheltered locations
were averaged and plotted for all times, there was a
consistent separation between assemblages in shel-
tered and exposed locations across all times (Fig. 2).
Because environmental conditions can affect patchi-
ness at multiple spatial scales, Bray-Curtis dissimilari-
ties estimating variation at the scale of metres (within
sites) and 100s of metres (between sites) were calcu-
lated for independent pairs of quadrats. Maintaining
their independence allows one to test the hypotheses
that these levels of variation are themselves inconsis-
tent among location and exposure using standard
ANOVA (Underwood & Chapman 1998a).
There was a significant interaction between times
and locations (F
36,112
= 1.91, p < 0.01) for analyses of the
Bray-Curtis dissimilarities between replicates within
sites (i.e. measuring variation at the scale of metres),
22
Roberts et al.: Sponge assemblages on temperate reefs
but there was no effect of exposure
(determined after pooling terms in the
analysis with p > 0.25; mean for exposed
locations = 49% dissimilarity, mean for
sheltered locations = 47% dissimilarity).
Variation at this scale was also tested
using the average Bray-Curtis dissimi-
larity per site and the sites as replicates,
and similar results were obtained (a sig-
nificant interaction between times and
locations; F
36,112
= 1.82, p < 0.01).
The Bray-Curtis dissimilarities mea-
sured between sites, i.e. variation in the
assemblage at the scale of 100s of
metres, similarly gave a significant
interaction between times and locations
(F
36,280
= 1.82, p < 0.01) and no general
effect of exposure after pooling (mean
for exposed locations = 52% dissimilar-
ity, mean for sheltered locations = 50%
dissimilarity).
Patterns in richness and cover
The cover of algae was generally
greatest on the exposed reefs, although
the number of species and their covers
fluctuated at various spatial and tempo-
ral scales (Fig. 3a to c, Table 1). The
cover of the silt matrix also fluctuated at
various spatial and temporal scales,
although it was not as variable through
time on the sheltered reefs as on the
exposed reefs (Fig. 3d, Table 1). The
total number of species and cover of all
the fauna combined fluctuated signifi-
cantly at the smallest scale examined
(Table 1), although both the number of
species and the cover of fauna were
generally greater on the sheltered reefs
(Fig. 3e,f).
There were significant spatial and
temporal fluctuations in the total num-
ber of species of sponge and cover of
sponges (Fig. 4a,b, Table 2). Generally,
greater richness and cover of sponges
were recorded on the sheltered reefs,
compared with the exposed reefs
(Fig. 4a,b). The richness and cover of
encrusting and erect sponges also
fluctuated through time (Fig. 4c to f,
Table 2). There were generally fewer
differences in the richness of encrust-
ing sponges between the 2 different
23
Apr 93
Stress = 0.15
Aug 93
Stress = 0.11
Jan 94
Stress = 0.17
Apr 94
Stress = 0.20
Jan 95
Stress = 0.18
Apr 95
Stress = 0.21
Sep 95
Stress = 0.14
All times averaged
Stress = 0.13
Sheltered reef
Exposed reef
Fig. 2. Non-metric multidimensional scaling plots for each time of sampling.
(Exposed : (
s) Bungan Head, (n) Long Reef, (e) North Head, (h) Cape
Banks. Sheltered : (
d) Quarantine Head, (m) South Head, (r) Henry Head,
(
j) Inscription Point)
Mar Ecol Prog Ser 321: 19–30, 200624
(a) Algae (b) Algae
(c) Crustose algae (d) Silt matrix
(e) Fauna (f) Fauna
MAMJJASOND JFMAMJJASOND JFMAMJJAS MAMJJASOND JFMAMJJASOND JFMAMJJAS
93 94 95 93 94 95
% Cover
40
30
20
10
0
% Cover
40
30
20
10
0
No. of Species
3
2
1
0
No. of Species
12
10
8
6
4
2
0
% Cover
60
40
20
0
% Cover
40
30
20
10
0
Fig. 3. Mean (±SE): (a) number of species of algae, (b) total cover of algae, (c) cover of crustose algae, (d) cover of silt matrix,
(e) number of species of fauna and (f) total cover of fauna on the reefs at: (
s) Bungan Head, (n) Long Reef, (e) North Head,
(
h) Cape Banks, (d) Quarantine Head, (m) South Head, (r) Henry Head and (j) Inscription Point. Total n = 15
Source df Algal richness Algal cover Crustose algae Silt matrix Faunal richness Faunal cover
MS F MS F MS F MS F MS F MS F
Time 6 11.4 256.1 658.7 2927.8 80.6 1875.2
Habitat 1 54.0 5181.8 960.8 2458.3 2506.4 43 935.3
Location (Habitat) 6 8.6 3922.6 4279.8 2878.0 163.3 3496.7
Site (Ti × Ha × Lo) 112 0.6 1.5** 123.3 1.4** 142.9 1.8** 363.3 1.4** 5.2 1.3* 224.0 1.3**
Ti × Ha 6 8.7 686.8 229.1 2332.5 3.9 350.2
Ti × Lo (Ha) 36 2.3 4.1** 426.5 3.5** 527.4 3.7** 1188.9 3.3** 16.9 3.3** 663.1 2.9**
Residual 672 0.4 85.9 78.1 252.4 3.9 169.2
Cochran’s test 0.058** 0.037, n.s. 0.036, n.s. 0.032, n.s. 0.029, n.s. 0.026, n.s.
Transformation None None Arcsine None None None
Table 1. Summaries of F-ratios from analyses comparing spatial and temporal variation in the species richness and cover of algae,
silt matrix and total fauna at locations in habitats on exposed and sheltered reefs in New South Wales, Australia (Ti: time; Ha:
habitat; Lo: location; None: no transformation required; n.s.: not significant [p > 0.05]; significant *p < 0.05, **p < 0.01)
Roberts et al.: Sponge assemblages on temperate reefs
25
(a) Sponges (b) Sponges
(c) Encrusting sponges (d) Encrusting sponges
(e) Erect sponges (f) Erect sponges
MAMJJASOND JFMAMJJASOND JFMAMJJAS MAMJJASOND JFMAMJJASOND JFMAMJJAS
93 94 95 93 94 95
% Cover
50
40
30
20
10
0
3
2
1
0
No. of SpeciesNo. of Species
10
8
6
4
2
0
No. of Species
7
6
5
4
3
2
1
0
% Cover
30
20
10
0
% Cover
40
30
20
10
0
Fig. 4. Mean (±SE): (a) number of species of sponges, (b) total cover of sponges, (c) number of species of encrusting sponges,
(d) cover of encrusting sponges, (e) number of species of erect sponges and (f) cover of erect sponges on the reefs at: (
s) Bungan
Head, (
n) Long Reef, (e) North Head, (h) Cape Banks, (d) Quarantine Head, (m) South Head, (r) Henry Head, and (j) Inscription
Point (n = 15)
Source df Sponge Encrusting Erect
richness cover richness cover richness cover
MS F MS F MS F MS F MS F MS F
Time 6 55.6 1027.4 15.2 217.3 18.4 143.8
Habitat 1 1190.5 39 373.8 16.3 2676.8 928.2 45 398.6
Location (Habitat) 6 104.4 3311.5 4.8 1262.8 75.3 1536.3
Site (Ti × Ha × Lo) 112 2.6 1.1, n.s. 182.7 1.1, n.s. 0.9 1.3** 78.1 0.9, n.s. 2.3 1.2, n.s. 99.5 1.3**
Ti × Ha 6 5.0 424.5 2.7 307.9 7.7 674.0
Ti × Lo (Ha) 36 8.7 3.3** 515.2 2.8** 2.1 2.2** 308.6 3.9** 5.6 2.5** 180.3 1.8**
Residual 672 2.3 162.5 0.7 80.3 1.8 79.0
Cochran’s test 0.029, n.s 0.034, n.s. 0.028, n.s. 0.055** 0.037, n.s. 0.039, n.s.
Transformation None None None None None Arcsine
Table 2. Summaries of F-ratios from analyses comparing spatial and temporal variation in the species richness and cover
of sponges at locations in habitats on exposed and sheltered reefs in New South Wales, Australia (Ti: time; Ha: habitat;
Lo: location; None: no transformation required; n.s.: not significant [p > 0.05]; significant *p < 0.05, **p < 0.01)
Mar Ecol Prog Ser 321: 19–30, 2006
habitats (Fig. 4c), whilst on some occasions, there
were greater covers of encrusting sponges on the
exposed reefs (Fig. 4d). The erect sponges were gen-
erally more diverse and had greater covers on the
sheltered reefs compared with those on the exposed
reefs (Fig. 4e,f).
The cover and richness of ascidians, bryozoans and
cnidarians fluctuated at various spatial and temporal
scales (Fig. 5a to f, Table 3). Their contribution to pri-
mary cover was very small in both habitats when com-
pared with the sponges. There were generally no pat-
terns in their species richness or cover between the
exposed or sheltered reefs, with the exception that, at
times, there were greater numbers of ascidian species
in some sheltered reef locations (Fig. 5a).
Examination of the components of variation for
each of the derived variables on exposed and shel-
tered reefs indicated that most of the variation was at
the smallest scale, i.e. within the residual mean
squares (Table 4). Relatively little variation was at the
scale of site or location, and there were many nega-
tive estimates at each of these scales (Table 4). Signif-
icant differences in the components of variation were
found in the residual (i.e. among replicates) using the
2-tailed F-tests (Table 4). Generally, there was signif-
icantly greater variability on exposed reefs than on
sheltered reefs for the cover of silt, algae and encrust-
ing sponges (Table 4). There was generally signifi-
cantly greater variability for the cover of erect
sponges on sheltered reefs (Table 4).
26
(a) Ascidians (b) Ascidians
(c) Bryozoans (d) Bryozoans
(e) Cnidarians (f) Cnidarians
MAMJJASOND JFMAMJJASOND JFMAMJJAS MAMJJASOND JFMAMJJASOND JFMAMJJAS
93 94 95 93 94 95
% Cover
8
6
4
2
0
1.5
1.0
0.5
0.0
No. of SpeciesNo. of Species
2
1
0
No. of Species
1.0
0.5
0.0
% Cover
4
3
2
1
0
% Cover
15
10
5
0
Fig. 5. Mean (±SE): (a) number of species of ascidians, (b) total cover of ascidians, (c) number of species of bryozoans, (d) cover of
bryozoans, (e) number of species of cnidarians and (f) cover of cnidarians on the reefs at: (
s) Bungan Head, (n) Long Reef, (e)
North Head, (
h) Cape Banks, (d) Quarantine Head, (m) South Head, (r) Henry Head, and (j) Inscription Point (n = 15)
Roberts et al.: Sponge assemblages on temperate reefs
DISCUSSION
Sponges were the dominant faunal group (greatest
richness and cover) on both exposed and sheltered
reefs, whilst ascidians were the next in importance. In
terms of the cover of primary space, foliose and crus-
tose macroalgae and a silt matrix were also important
in contributing to the structure of the assemblages. The
silt matrix was generally responsible for the greatest
primary cover on the exposed reefs, whilst sponges
had the greatest cover on sheltered reefs. The cover of
macroalgae (foliose and crustose) was generally great-
est on exposed reefs, whereas the covers of the other
fauna (ascidians, bryozoans and cnidarians) were gen-
erally similar in terms of the amount of space they
occupied in both habitats.
The greater richness and cover of sponges on shel-
tered reefs was primarily due to the presence of more
massive or erect forms, e.g. Spirastrella sp., Desmap-
samma kirki Bowerbank and Mycale spp. Whilst these
species were also found on the exposed reefs, their
morphology or form of growth tended to be more pros-
trate. Patterns of increased cover and richness of pros-
trate sponges with decreasing depth have been
described for temperate (Roberts & Davis 1996, Bell &
Barnes 2000) and tropical reefs (Wilkinson & Evans
1989). Energy generally decreases with depth, which
may partly explain these types of patterns. In general,
those species with small basal area to volume ratios do
poorly in high-energy environments (Wulff 1995, Bell
& Barnes 2000). Greater richness and cover of encrust-
ing sponges were found on the exposed reefs, where
there was generally more water turbulence. Depth-
related patterns associated with turbulence have also
been described for other types of subtidal assemblages
(Schmahl 1990, Clarke et al. 1993, Bell & Barnes 2000).
Increased sedimentation, light and water turbulence
associated with depth gradients have all been identi-
fied as important factors in structuring the distribution
and abundance of subtidal assemblages of sponges in
temperate (Underwood et al. 1991, Zea 1993, Roberts
& Davis 1996, Bell & Barnes 2000, Hooper & Kennedy
2002), tropical (Liddell & Ohlhorst 1987, Wilkinson &
Evans 1989) and polar (Barthel 1991) regions. In the
present study we can exclude patterns of distribution
and abundance of sponges being related to depth gra-
dients, because consistently different community
structures were found between the 2 habitat types, at
the same depth. Underwood et al. (1991) described a
mosaic of subtidal habitat types along the coastline of
New South Wales, Australia, with the distribution of
assemblages related to depth, wave exposure and her-
bivory, although their study did not extend into the
habitats below 15 m.
Habitat-related differences have also been de-
scribed at smaller spatial scales (Wright et al. 1997),
where differences in the structure of sponge assem-
blages between 2 adjacent habitats at the same depth
were related to the ability of some sponges to with-
stand predation by using chemical defences. In shal-
low subtidal habitats, grazing sea urchins and molluscs
have been found to effectively maintain the structure
of coralline algal crusts in barren ground habitats
(Davis et al. 2003). At this stage, however, the pro-
cesses that interact to determine the structure of subti-
dal assemblages on reefs at depths >18 to 20 m are
largely unknown (Underwood et al. 1991).
In general, the analyses of components of variation
supported the hypothesis of increased variability for
the assemblages on the exposed reefs. Warwick &
Clarke (1993) identified increased variability in
meiobenthic, macrobenthic, coral reef and fish assem-
blages associated with different types of disturbance in
natural habitats, and concluded that this variability
27
Source df Ascidian Bryozoan Cnidarian
richness cover richness cover richness cover
MS F MS F MS F MS F MS F MS F
Time 6 6.6 191.8 0.8 1.6 1.4 97.9
Habitat 1 65.2 1472.1 15.2 24.7 0.8 167.9
Location (Habitat) 6 10.1 489.8 1.6 4.3 2.7 332.4
Site (Ti × Ha × Lo) 112 0.9 1.4** 35.7 1.2, n.s. 0.3 0.9, n.s. 0.9 0.9, n.s. 0.3 1.5** 27.1 1.8**
Ti × Ha 6 0.8 21.7 1.1 3.9 0.9 69.9
Ti × Lo (Ha) 36 1.3 1.5* 55.6 1.6* 0.6 1.9** 2.9 3.1** 0.6 1.8** 65.7 2.4**
Residual 672 0.6 29.7 0.3 1.0 0.2 15.2
Cochran’s test 0.028, n.s. 0.035, n.s. 0.029, n.s. 0.213** 0.031, n.s. 0.212**
Transformation None Arcsine None None None None
Table 3. Summaries of F-ratios from analyses comparing spatial and temporal variation in the species richness and cover of ascid-
ians, bryozoans and cnidarians at locations in habitats on exposed and sheltered reefs in New South Wales, Australia (Ti: time;
Ha: habitat; Lo: location; None: no transformation required; n.s.: not significant [p > 0.05]; significant *p < 0.05, **p < 0.01)
Mar Ecol Prog Ser 321: 19–30, 2006
may be a symptom of stress (but see Chapman et al.
1995). Wave energy and storms are considered to be
natural physical disturbances in temperate subtidal
reefs within embayments, estuaries and on exposed
coastlines (Kennelly 1989, Underwood 1999). The
greatest changes in assemblages will occur where the
disturbance is not one that the assemblage normally
experiences and will depend on its type and magni-
tude, the pre-disturbance structure of the community
(Roberts et al. 1998) and the morphological and physi-
ological adaptations of its members (Schratzberger &
Warwick 1999).
There were significant interactions at small spatial
scales for many of the variables we examined. These
interactions are ecologically important and show how
assemblages experience patchiness in their distribu-
tions at different spatial scales. The incorporation of a
hierarchy of spatial scales into sampling programmes,
as well as the need for appropriate spatial scales
required to detect the effects of anthropogenic distur-
bance, was highlighted by Underwood (2000). Signifi-
cant variability was found at all the spatial scales we
examined, although the variability among replicate
quadrats consistently contributed to the greatest pro-
portion of the total variation. This reinforces the model
that patchiness at small spatial scales, which is caused
by local processes, is important in structuring assem-
blages rather than processes that operate at larger
scales. Local processes and small-scale heterogeneity
within a site may cause increased variability that is far
greater than what may be caused by processes at
larger scales. From a managerial perspective, if small
scales are not included in sampling programmes that
measure environmental disturbances, then impacts
may go undetected (Underwood et al. 2003).
If we ignore for the moment that assemblages of sub-
tidal marine organisms living on coastal reefs close to
large cities may be affected by various forms of anthro-
pogenic disturbance, then natural physical processes,
such as storms, may account for a large proportion of
the variation in differences between habitats (Under-
wood et al. 1991). Physical, chemical and biological
factors, which could potentially determine the struc-
ture and dynamics of assemblages on these subtidal
reefs, would need to be tested using appropriate
manipulative experiments. The potential physical dis-
tinctions between the habitats studied here include
increased wave energy (Short & Trenaman 1992) and
light penetration on exposed reefs (Rendell & Pritchard
1996), with stronger tidal currents and increased silta-
tion on the sheltered estuarine reefs (Middleton et al.
1996). There would also be greater loads of nutrients
and fluctuations in salinity on the reefs in the estuarine
locations (Middleton et al. 1996, Rendell & Pritchard
1996). Differences related to predation (Ayling 1981),
28
Taxa Treatment Time 1 Time 2 Time 3 Time 4 Time 5 Time 6 Time 7
Re Si Lo Re Si Lo Re Si Lo Re Si Lo Re Si Lo Re Si Lo Re Si Lo
Silt Exp 337.3 3.5 2.9 208.9 3.7 11.6 314.9 – 31.3 226.9 11.5 – 412.3 7.6 – 409.5 – 51.9 284.1 – –
matrix She 232.7 3.7 2.7 152.7 0.2 0.9 123.1 – 13.2 166.1 1.5 – 238.0 0.2 – 230.9 – 36.8 206.4 – –
F-test 1.5* 1.4* 2.6* 1.4* 1.7* 1.8* 1.4*
Total Exp 228.5 0.9 – 213.5 0.99 6.7 154.4 – 3.3 85.5 7.5 2.1 102.0 – 1.0 154.1 – 56.0 193.9 – 13.8
algae She 84.9 0.8 – 63.8 0.06 1.3 71.4 – 20.6 98.1 3.5 1.6 64.6 – 7.0 69.9 – 18.7 207.6 – 22.4
F-test 2.7* 3.4* 2.2* 1.2* 1.6* 2.2* 1.0, n.s.
Total Exp 109.0 – 2.9 129.8 – 13.4 131.1 1.4 8.6 245.6 – 24.7 257.1 – 10.1 176.2 – 9.2 145.6 1.0 2.3
fauna She 105.2 – 15.4 180.4 – 35.5 156.4 6.9 0.6 157.9 – 5.6 176.3 – 1.1 239.8 – 17.3 169.4 4.9 0.3
F-test 1.0, n.s. 1.4* 1.2* 1.6* 1.5* 1.4* 1.2*
Total Exp 106.1 – 3.5 152.7 – 15.0 136.6 0.6 8.1 132.3 – 5.2 219.6 – 5.8 182.1 – 3.4 143.7 0.3 –
sponges She 106.2 – 18.5 115.5 – 11.9 142.7 7.9 1.3 200.4 – 16.5 162.5 – 5.6 264.3 – 12.2 172.6 8.1 –
F-test 1.0, n.s. 1.3* 1.0, n.s. 1.5* 1.4* 1.5* 1.2*
Encrust. Exp 38.2 – 5.3 98.1 – 32.9 135.6 – 20.9 170.9 – 8.5 109.5 1.2 0.5 163.4 – 2.3 102.6 – 2.7
sponges She 41.2 – 2.2 48.1 – 2.9 56.1 – 1.3 69.9 – 5.5 43.5 0.5 0.2 11.5 – 0.8 19.5 – 0.2
F-test 1.1* 2.0* 2.4* 2.4* 2.5* 14.2* 5.3*
Erect Exp 99.8 – 0.2 78.7 – 10.5 67.3 0.05 1.2 46.6 – 0.9 186.2 0.8 – 70.1 0.6 0.12 80.4 – 2.6
sponges She 96.8 – 6.7 48.1 – 2.9 197.8 4.8 1.2 174.5 – 3.3 154.9 0.04 – 254.6 3.8 1.12 188.9 – 23.2
F-test 1.0, n.s. 1.6* 2.9* 3.7* 1.2* 3.6* 2.4*
Table 4. Estimates of components of variation derived from the analyses of variance between exposed and sheltered reefs for cover of selected taxa at each time
(n.s.: not significant [p > 0.05]; significant *p < 0.05; Exp: exposed; She: sheltered; Re: residual; Si: site; Lo: location)
Roberts et al.: Sponge assemblages on temperate reefs
recruitment (Butler 1986) and competition for space
(Ayling 1983) cannot be discounted as potentially
important determinants of the structure and dynamics
of assemblages within these 2 types of habitat. As
Underwood (2000) noted and Hill & Hill (2002) demon-
strated, hypotheses derived from models such as those
outlined above must be tested by manipulation of the
regime of disturbances.
Presenting these results was an important, logical,
first-step in developing an understanding of the ecol-
ogy of sponge-dominated assemblages on subtidal
reefs at depths of 18 to 20 m. Research into the ecology
of subtidal assemblages living on hard substrata has
unfortunately been ‘depth limited’, with most effort
spent on those shallow-water assemblages, which are
easily sampled using SCUBA techniques (Underwood
et al. 1991, Davis & Ward 1999) or by examining
assemblages on artificial structures (Kay & Butler 1983,
Glasby 1999). In recent years, marine ecologists have
rightfully focussed on experimental tests of hypotheses
about processes to explain patterns of variability; nev-
ertheless, descriptive quantitative tests of mensurative
models are still a necessary pre-cursor to experimental
manipulative analyses (Underwood et al. 2000). For
sponge-dominated assemblages on deeper temperate
reefs, experimental tests of hypotheses about the pro-
cesses producing their structure and dynamics were
considered to be premature, until spatial and temporal
variability has been quantified at appropriate scales.
Acknowledgements. We thank J. Hooper at the Queensland
Museum for comments and identification of the sponge fauna
and B. Roberts for assistance in the field. R. Reinfrank (EICC)
is thanked for assistance with some of the multivariate analy-
ses. D. Leece, T. Church, P. Scanes and K. Koop (Department
of Environment and Conservation) provided logistic support
for the study, which represents Contribution No. 273 from the
Ecology & Genetics Group, University of Wollongong.
LITERATURE CITED
Anderson MJ (2001) A new method for non-parametric multi-
variate analysis of variance. Aust Ecol 26:32–46
Archambault P, Bourget E (1996) Scales of coastal hetero-
geneity and benthic intertidal species richness, diversity
and abundance. Mar Ecol Prog Ser 136:111–121
Ayling AL (1983) Factors affecting the spatial distributions of
thinly encrusting sponges from temperate waters. Oecolo-
gia 60:412–418
Ayling AM (1981) The role of biological disturbance in tem-
perate subtidal encrusting communities. Ecology 62:
830–847
Barthel D (1991) Influence of different current regimes on the
growth form of Halichondria panicea, Pallas. In: Reitner J,
Keupp H (eds) Fossils and recent sponges. Springer-
Verlag, Berlin
Bell JJ, Barnes KA (2000) The influences of bathymetry and
flow regime upon the morphology of sublittoral sponge
communities. J Mar Biol Assoc UK 80:707–718
Bishop MJ, Underwood AJ, Archambault P (2002) Sewage
and environmental impacts on rocky shores: necessity of
identifying relevant spatial scales. Mar Ecol Prog Ser 236:
121–128
Butler AJ (1986) Recruitment of sessile invertebrates at five
sites in Gulf St Vincent, South Australia. J Exp Mar Biol
Ecol 97:13–36
Butler AJ, Connolly RM (1996) Development and long term
dynamics of a fouling assemblage of sessile marine inver-
tebrates. Biofouling 9:187–209
Carballo JL, Naranjo SA, Garcia-Gomez JC (1996) The use of
sponges as stress indicators in marine ecosystems at Alge-
ciras Bay (southern Iberian Peninsula). Mar Ecol Prog Ser
135:109–122
Chapman MG, Underwood AJ, Skilleter GA (1995) Variabil-
ity at different spatial scales between a subtidal assem-
blage exposed to the discharge of sewage and two control
assemblages. J Exp Mar Biol Ecol 189:103–122
Clarke KR, Warwick RM, Brown BE (1993) An index showing
breakdown of seriation, related to disturbance, in a coral-
reef assemblage. Mar Ecol Prog Ser 102:153–160
Coleman MA (2002) Small-scale spatial variability in inter-
tidal and subtidal turfing algal assemblages and the tem-
poral generality of these patterns. J Exp Mar Biol Ecol 267:
53–74
Davis AR, Ward DW (1999) Does the large barnacle Aus-
trobalanus imperator (Darwin, 1854) structure benthic
invertebrate communities in SE Australia? Mem Qld Mus
44:125–130
Davis AR, Fyfe SK, Turon X, Uriz MJ (2003) Size matters
sometimes: wall height and the structure of subtidal ben-
thic invertebrate assemblages in southeastern Australia
and Mediterranean Spain. J Biogeogr 30:1797–1807
Fletcher DJ, Underwood AJ (2002) How to cope with negative
estimates of components of variation in ecological field
studies. J Exp Mar Biol Ecol 273:89–95
Ginn BK, Logan A, Thomas MLH (2000) Sponge ecology on
sublittoral hard substrates in a high current velocity area.
Estuar Coast Shelf Sci 50:403–414
Glasby TM (1999) Differences between subtidal epibiota on
pier pilings and rocky reefs at marinas in Sydney, Aus-
tralia. Estuar Coast Shelf Sci 48:281–290
Graham MH, Harrold C, Lisin S, Light K, Watanabe JM, Fos-
ter MS (1997) Population dynamics of giant kelp Macro-
cystis pyrifera along a wave exposure gradient. Mar Ecol
Prog Ser 148:269–279
Hewitt JE, Thrush SF, Legendre P, Cummings VJ, Norkko A
(2002) Integrating heterogeneity across spatial scales:
interactions between Atrina zelandica and benthic macro-
fauna. Mar Ecol Prog Ser 239:115–128
Hill MS, Hill AL (2002) Morphological plasticity in the tropical
sponge Anthosigmella varians: responses to predators and
wave energy. Biol Bull (Woods Hole) 202:86–95
Hooper JNA, Kennedy JA (2002) Small-scale patterns of bio-
diversity in sponges (Porifera), from the Sunshine Coast,
southeast Queensland. Invertebr Syst 16:637–653
Hummel H, Sepers ABJ, de Wolf L, Melissen FW (1988) Bac-
terial growth on the marine sponge Halichondria panicea
induced by reduced waterflow rate. Mar Ecol Prog Ser 42:
195–198
Kay AM, Butler AJ (1983) ‘Stability’ of the fouling communi-
ties on the pilings of two piers in South Australia. Oecolo-
gia 56:58–66
Kennelly SJ (1989) Effects of kelp canopies on understorey
species due to shade and scour. Mar Ecol Prog Ser 50:
215–224
29
Mar Ecol Prog Ser 321: 19–30, 2006
Liddell WD, Ohlhorst SL (1987) Patterns of reef community
structure, North Jamaica. Bull Mar Sci 40:311–329
Maldonado M, Uriz MJ (1998) Microrefuge exploitation by
subtidal encrusting sponges: patterns of settlement and
post-settlement survival. Mar Ecol Prog Ser 174:141–150
Maldonado M, Young CM (1996) Effects of physical factors on
larval behaviour, settlement and recruitment of four tropi-
cal demosponges. Mar Ecol Prog Ser 138:169–180
McDonald JI, Hooper JNA, McGuinness KA (2002) Environ-
mentally influenced variability in the morphology of
Cinachyrella australiensis (Carter, 1886) (Porifera:
Spirophorida: Tetillidae). Mar Freshw Res 53:79–84
McDonald JI, McGuinness KA, Hooper JNA (2003) Influence
of re-orientation on alignment to flow and tissue produc-
tion in a Spongia sp. (Porifera: Demospongiae: Dictyocer-
atida). J Exp Mar Biol Ecol 296:13–22
Middleton JH, Cox D, Tate P (1996) The oceanography of the
Sydney region. Mar Pollut Bull 33:124–131
Palumbi SR (1986) How body plans limit acclimation: responses
of a demosponge to wave force. Ecology 67:208–214
Posey M, Lindberg W, Alphin T, Vose F (1996) Influence of
storm disturbance on an offshore benthic community. Bull
Mar Sci 59:523–529
Rendell PS, Pritchard TR (1996) Physicochemical conditions in
coastal waters off Sydney, Central NSW, Australia. Mar
Pollut Bull 33:132–139
Roberts DE, Davis AR (1996) Patterns in sponge (Porifera)
assemblages on temperate coastal reefs off Sydney, Aus-
tralia. Mar Freshw Res 47:897–906
Roberts DE, Smith A, Ajani P, Davis AR (1998) Rapid changes
in encrusting marine assemblages exposed to anthro-
pogenic point-source pollution: a ‘Beyond BACI’
approach. Mar Ecol Prog Ser 163:213–224
Schmahl GP (1990) Community structure and ecology of
sponges associated with four southern Florida coral reefs.
In: Rutzler K (ed) New perspectives in sponge biology.
Smithsonian Institute Press, London
Schratzberger M, Warwick RM (1999) Differential effects of
various types of disturbance on the structure of nematode
assemblages: an experimental approach. Mar Ecol Prog
Ser 181:227–236
Short AD, Trenaman NL (1992) Wave climate of the Sydney
region, an energetic and highly variable ocean wave
regime. Aust J Mar Freshw Res 43:765–791
Stewart-Oaten A, Bence JR (2001) Temporal and spatial vari-
ation in environmental impact assessment. Ecol Monogr
71:305–339
Thompson RC, Wilson BJ, Tobin ML, Hill AS, Hawkins SJ
(1996) Biologically generated habitat provision and diver-
sity of rocky shore organisms at a hierarchy of spatial
scales. J Exp Mar Biol Ecol 202:73–84
Underwood AJ (1997) Experiments in ecology: their logical
design and interpretation using analysis of variance. Cam-
bridge University Press, Cambridge
Underwood AJ (1998) Grazing and disturbance: an experi-
mental analysis of patchiness in recovery from a severe
storm by the intertidal alga Hormosira banksii on rocky
shores in New South Wales. J Exp Mar Biol Ecol 231:
291–306
Underwood AJ (1999) Physical disturbances and their effect
on an indirect effect: responses of an intertidal assem-
blage to a severe storm. J Exp Mar Biol Ecol 232:125–140
Underwood AJ (2000) Importance of experimental design in
detecting and measuring stress in marine populations.
J Aquat Ecosyst Stress Recovery 7:3–24
Underwood AJ, Chapman MG (1998a) A method for
analysing spatial scales of variation in composition of
assemblages. Oecologia 117:570–578
Underwood AJ, Chapman MG (1998b) Variation in algal
assemblages on wave-exposed rocky shores in New South
Wales. Mar Freshw Res 49:241–254
Underwood AJ, Kingsford MJ, Andrew NL (1991) Patterns in
shallow subtidal marine assemblages along the coast of
New South Wales. Aust J Ecol 6:231–249
Underwood AJ, Chapman MG, Connell SD (2000) Observa-
tions in ecology: you can’t make progress on processes
without understanding the patterns. J Exp Mar Biol Ecol
250:97–115
Underwood AJ, Chapman MG, Roberts DE (2003) A practical
protocol to assess impacts of unplanned disturbance: a
case study in Tuggerah Lakes estuary, NSW. Ecol Manag
Restor 4:4–11
Warwick RM, Clarke KR (1993) Increased variability as a
symptom of stress in marine communities. J Exp Mar Biol
Ecol 172:215–226
Wilkinson CR, Evans E (1989) Sponge distribution across
Davies Reef, Great Barrier Reef, relative to location,
depth, and water movement. Coral Reefs 8:1–7
Winer BJ (1971) Statistical principles in experimental design.
McGraw-Hill, New York
Wright JT, Benkendorff K, Davis AR (1997) Habitat associated
differences in temperate sponge assemblages: the impor-
tance of chemical defence. J Exp Mar Biol Ecol 213:
199–213
Wulff JL (1995) Effects of a hurricane on the survival and ori-
entation of large erect coral reef sponges. Coral Reefs 14:
55–61
Zea S (1993) Cover of sponges and other sessile organisms in
rocky and coastal reef habitats of Santa Marta, Colombian
Caribbean Sea. Caribb J Sci 29:75–88
30
Editorial responsibility: Roger N. Hughes (Contributing
Editor), Bangor, UK
Submitted: September 20, 2005; Accepted: January 25, 2006
Proofs received from author(s): August 22, 2006
